Read the original
Prefer a site's native Markdown, source APIs, and high-quality extraction. Planned for read_link.
Origo / origin, source
A second path to the source.
An MCP server built for moments when an agent can find a page but cannot properly read it. Origo will focus on direct link retrieval and site mapping, not search.
Infrastructure preview. The server transport and authentication are scaffolded; retrieval features are still being designed.
Design principles
Optimized for reading. Not another search engine.
Origo is intended to preserve the strengths of WebCTX's Markdown and GitHub retrieval while giving agents more ways to access pages when a standard browser fetch falls short.
Explore a site
Discover relevant URLs without reimplementing web search. Planned for map_site.
Fallback only
Designed to complement native agent browsing and search, not replace either.
Small surface
A stateless Go MCP endpoint on Vercel with API-key authentication. Infrastructure scaffold is ready.
Origo is a private, early-stage project. The public docs describe the architecture, not a finished retrieval product.